Core Concepts
The core message of this article is to evaluate the feasibility of leveraging Multi-Access Edge Computing (MEC) to support the computational requirements of various Connected and Automated Vehicle (CAV) services, such as remote driving, cooperative sensing, cooperative maneuver, and cooperative awareness, in a 5G network environment.
Abstract
The article presents a simulation-based study to assess the computational capacity that MEC nodes must be equipped with in order to support different CAV services. It first provides a basic queueing theory model to determine the minimum computational resources required for each service to meet the stringent latency and reliability requirements.
The authors then conduct a comprehensive simulation campaign using various open-source tools, including OMNeT++, Simu5G, Veins, INET, and SUMO, to simulate realistic vehicle mobility, 5G network conditions, and the communication and computation behavior of the CAV services.
The key findings are:
For some CAV services like remote driving and cooperative sensing, the required MEC resources are high, and only a few vehicles can be supported by a single MEC node, making MEC a challenging approach.
For other services like cooperative maneuver and awareness, MEC is more promising, as it can support a larger number of vehicles.
The computational capacity required varies significantly across the different CAV services, highlighting the need for careful planning and dimensioning of MEC deployments to support the diverse requirements.
In addition to the MEC node's computational capacity, the communication delays also play a crucial role in the feasibility of edge-based control of CAVs, which needs to be further investigated.
The study provides a valuable reference for network operators to plan the future deployment of MEC infrastructure to support CAV services.
Stats
The minimum CPU that should be allocated to each MecApp in order to respect the necessary condition for meeting the reliability requirement is:
Remote driving: 165130 MIPS
Cooperative sensing: 79915 MIPS
Cooperative maneuver: 28026 MIPS
Cooperative awareness: 7992 MIPS